Question

I founds this error in multiple question but it is not fixed or explained
Ex. Reading in single str column with loadtxt

I have a problem with the following line (import numpy as np)

    symbols = np.loadtxt('filename.csv',
               dtype={'names': ('ticker', 'year', 'month','day'),
                      'formats': ('S10','i4','i2','i2')},
                delimiter=',',skiprows=1, usecols=(0,))

with following data

ticker  year    month   day
GPD.TO  2010    8   16
HAO.V   2010    8   16
RDS.V   2010    8   16
MD.V    2010    8   16

It gives me a

IndexError: list index out of range

When I run that line without usecols it works fine and returns all 4 columns fine.

I looked at many other questions on google and Stackoverflow but this isnt answered

Answer 1

If you want to use usecols = (0,) with np.loadtxt then you must change the dtype to match:

import numpy as np
symbols = np.loadtxt('filename.csv',
                     dtype={'names': ('ticker', ),
                            'formats': ('S10',)},
                     delimiter=',',
                     skiprows=1,
                     usecols=(0,)
                     )
print(symbols)

with filename.csv:

ticker,year,month,day
GPD.TO,2010,8,16
HAO.V,2010,8,16
RDS.V,2010,8,16
MD.V,2010,8,16

yields

[('GPD.TO',) ('HAO.V',) ('RDS.V',) ('MD.V',)]

---

Alternatively, with np.genfromtxt you do not need to change dtype:

symbols = np.genfromtxt('filename.csv',
                        dtype={'names': ('ticker', 'year', 'month','day'),
                               'formats': ('S10','i4','i2','i2')},                        
                        delimiter=',',
                        skiprows=1,
                        usecols=(0,)
                        )

works as well.